For example, there are numbers that are incredibly unlikely, verging on impossible. In table 4, the study shows mean, standard deviation, and ranges for recovery time in patients within the study. The issue is that with a reported range of 9–25 days, a mean of 18 and a standard deviation of 8 there are very few configurations of numbers that would leave you with this result. You can even calculate this yourself
using the SPRITE tool developed by the clever fraud detectives James Heathers, Nick Brown, Jordan Anaya, and Tim Van Der Zee — to have a mean of 18 days consistent with the other values, the majority of the patients in this group would have to have stayed in the hospital for either 9 or 25 days
exactly. Now that might not be entirely impossible, but it’s so odd that it raises very serious questions about the results of the trial itself.